The prediction and explanation of longer-term functioning following the treatment of alcohol use disorders (AUD) have clinical, research, and theoretical importance. The overwhelming majority of alcohol treatment outcome research has been limited to pre- and early treatment predictors of longer-term functioning. To date, there has been relatively little research on the relationship of behaviors during or initially following an episode of alcohol treatment and later functioning. This question is critical because of its relevance to behavior change and its maintenance, which is fundamental to the success of interventions for behavioral or psychiatric disorders. The research that is available shows an association between early post-treatment drinking behavior, particularly abstinence, and later functioning, but the mechanisms underlying this relationship are not known. The long-term goals of this research are to enhance the power of alcohol use disorders (AUD) interventions to facilitate improvement in substance use and other functioning, and to sustain such changes. The proposed research involves secondary data analyses of Project MATCH data sets (i.e., outpatient study arm three-year outcome data, medical care costs study data, and the Albuquerque, NM clinical research unit 10-year outcome data) to achieve the following specific aims: 1) to determine the relationship between sustained short-term (up to 1 year) abstinence from alcohol following the initiation of alcohol treatment and longer-term substance use and other areas of functioning (e.g., marital, psychological, social); 2) to determine whether a period of short-term, moderate (characterized by no "heavy" drinking) drinking predicts longer-term functioning; and 3) to evaluate models of the mediation or moderation of psychosocial factors (e.g., self-efficacy, social/family support, psychological functioning) that underlie the association between shorter-term alcohol use and longer-term functioning. The results of this research will provide the basis for a subsequent standard research application (R01) aimed at longitudinally testing the models derived from this R21 application.